The kinematic analysis of motion curves through MIDI data analysis

  • Authors:

  • Mitali Das;David M. Howard;Stephen L. Smith

  • Affiliations:

  • Electronics Department, University of York, Heslington, York YO10 5DD, UK md123@ohm.york.ac.uk Fax: +44 (1904) 432335;Electronics Department, University of York, Heslington, York YO10 5DD, UK dmh@ohm.york.ac.uk Fax: +44 (1904) 432335;Electronics Department, University of York, Heslington, York YO10 5DD, UK sls@ohm.york.ac.uk Fax: +44 (1904) 432335

  • Venue:
  • Organised Sound
  • Year:
  • 1999

  • Lyrics, music, and emotions

    EMNLP-CoNLL '12 Proceedings of the 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper postulates that the analysis of MIDI data allows for the statistical analysis of motion in music, in particular, the kinematic analysis of motion curves in music. The paper deals with the analysis of the kinematic motion components within music, specifically music velocity (i.e. tempo) and music acceleration/deceleration (i.e. tempo change), based upon Truslit's definition of the predominant up–down motion types in music (1938). Thus, the variables of music velocity and acceleration are mathematically defined and extracted from MIDI encodings. Analysis of music velocity indicates that the differing motion types have specific and consistent velocity profiles, and that these profiles can be expressed mathematically and analysed statistically. In particular, the paper focuses on the open motion curve fit, relating the open motion velocity curve to the beta distribution. Analysis of acceleration within music suggests that music acceleration is not constant in nature, implying that theories of linear velocity are inaccurate models. Hence, it is suggested that MIDI data analysis allows for the statistical exploration into music kinematics and the motion curves within music.